Astrocyte is the most abundant glia cell and significantly outnumbers neuron in the human brain. Long thought to be primarily passive cell, astrocyte has been increasingly recognized as essential player with active regulatory role in neural circuitry and pathology. Since a single astrocyte interacts with thousands of synapses, other glial cells and blood vessels, it is well positioned to link neuronal information in different spatial-temporal dimensions to achieve higher level brain integration. Indeed, neuron-astrocyte communication at synapses regulates breathing, memory formation, motor function, and sleep, and are implicated in many neuropsychiatric disorders. All these results provide strong rationale for modeling and analyzing astrocyte function, which will provide unprecedented insights to our understanding how astrocytes function to regulate and protect brain and how these functions can be exploited for astrocyte-based therapeutic targets. Recent advances in the modern microscopy and ultrasensitive genetic encoded calcium indicators (GECI) have enabled optical recording of astrocytic calcium dynamics ? the excitatory state and functional readout of astrocytes ? in vitro, ex vivo and in vivo. Compared to the great experimental capability of generating tremendous volumes of astrocyte functional data, the development of computational tools to analyze and interpret the complex and big data is lagged far behind, which has severely jeopardized a deeper understanding of the functional roles of astrocytes. To address the pressing need, we thus propose to develop sophisticated computational tools for interpreting the complex calcium dynamics data, through judicious application of advanced machine learning and systems theories. We have the following three specific aim. Aim1). Developing computational tools to analyze the cellular properties of calcium signaling in a single astrocyte. Aim2). Developing computational tools to analyze the network properties of calcium signaling in a population of astrocytes. Aim3). Validating experimentally the computational tools, developing optimal experiment protocol and disseminating the software packages. Our preliminary studies on both synthetic and real datasets demonstrate the feasibility of our plans and highlight the potential of analyzing astrocyte functional activity to understand neuronal circuitry and pathology, including for the first time the surprising discovery of hyper-activity in Down?s syndrome astrocytes compared to the normal astrocytes. This proposal is built on pre-established collaboration between two groups with the much needed complementary expertise for accomplishing this project, (1) computational scientists (Yu lab at Virginia Tech) and (2) experimental neuroscientists (Tian lab at UC Davis). The pre-established working relationships, developed channels of communication and mechanisms for resource sharing will help insure that the work will proceed in an efficient and effective manner.